Image source: NVIDIA.

NVIDIA (NVDA -1.32%) is known for its leadership in the graphics processing units (GPUs) business, but the company is also helping to power computers for some of the world's biggest technology companies too.

Facebook (META -0.62%), Microsoft (MSFT 0.06%), and (AMZN -0.22%) are increasingly looking to GPUs to power the most complex parts of their cloud computing and deep learning technology -- and NVIDIA could benefit in the long term. 

How Microsoft's teaming up with NVIDIA

The Windows-maker has tapped NVIDIA's GPU prowess for a handful of its devices, including a custom GPU for the Surface Book and NVIDIA's Tegra processors in earlier versions of its Surface tablets.

But Microsoft upped its reliance on GPUs when it decided to partner with NVIDIA last year to bring the company's GPUs into its Azure cloud platform. 

The deal brought NVIDIA's GPUs, Grid 2.0 virtualized graphics, and Tesla platform to Azure. This allows Microsoft's cloud network to bring professional-grade graphics processing and super-computing performance to computers that run on the Azure cloud. 

Microsoft is expanding its role in the increasingly competitive cloud computing market, and adding the company's GPUs makes its Azure cloud network more competitive with Amazon's offerings. 

What Amazon's doing with NVIDIA

Amazon's Web Service (AWS) has been using NVIDIA's GPUs to power graphics processing on its cloud service since at least 2013, and it's very similar to how Microsoft is using NVIDIA's GPU's for Azure.

Amazon's AWS website says GPUs are "ideally suited for graphical high performance computing applications" for 3D modeling, financial analysis, engineering design and other applications. 

Amazon's use of NVIDIA's GPUs is notable because of the company's position in the cloud computing market. Amazon brought in $2.5 billion from AWS in just the first quarter of 2016 alone, up 63% year-over-year. And as Amazon's AWS grows, sales of NVIDIAs processors could certainly benefit. 

How Facebook's using NVIDIA's tech

Facebook's use of NVIDIA came to light recently when NVIDIA said the social media company is using its Tesla M40 GPUs for its Big Sur machine learning computers. NVIDA says the neural network GPU platform used by Facebook is 10 to 20 times faster than computers with similar specifications running on traditional central processing units (CPUs). 

Facebook's use of NVIDIA's GPUs is notable because the machine learning space -- where computers learn things they weren't preprogrammed to know -- is growing quickly.

The broader cognitive computing market (which includes machine learning) will be worth an estimated $12 billion by 2019, according to MarketsandMarkets. And GPUs are at the heart of these systems. NVIDIA say its GPUs can power neural network computers to learn new things in a few hours, compared to days or weeks with non-GPU computers. 

Great potential later

NVIDIA may be expanding its presence in the cloud computing and machine learning space, but investors should remember that revenues from these segments still makes up a relatively small portion of the company's total revenue. 

In fiscal Q1 2017, NVIDIA made just $143 million from its datacenter segment -- meanwhile, the company's main revenue driver, its gaming segment, brought in $687 million.

NVIDIA is certainly adding value to its business by providing deep learning and cloud-computing GPUs to Facebook, Microsoft, and Amazon, but don't expect datacenter revenue to overtake gaming any time soon.

It's best to view NVIDIA's GPU supplies to major tech companies as an inroad to bigger cloud-based offerings over the next few years. I wouldn't invest in the company solely based on its movements in the space right now, but they should pay off for the company in the coming years.